Only one of the major ERP support deadlines is both close and firm. SAP ends mainstream maintenance for ECC on December 31, 2027, with paid extended maintenance to 2030 and a private cloud transition option to 2033. Oracle has committed Premier Support for E-Business Suite, JD Edwards, and PeopleSoft through at least 2037. Microsoft supports Dynamics GP to the end of 2029, with security patches into 2031, and Dynamics AX has been out of support since January 2023. A tier 1 ERP program takes about 36 months from decision to go-live, so the date on the vendor's lifecycle page is less important than the date 36 months before it. That is the one most companies have already passed.
The deadlines
| Vendor and product | Mainstream support ends | Paid extension ends | What you can still get after mainstream ends | Source |
|---|---|---|---|---|
| SAP ERP 6.0 (ECC), enhancement packs 6 to 8 | December 31, 2027 | December 31, 2030 (extended maintenance); 2033 (private cloud transition option, sign before 2028) | Extended maintenance keeps close to mainstream scope at a premium; after that, customer-specific maintenance fixes known problems only | Rimini Street, ERP Today |
| SAP ERP 6.0, enhancement packs 0 to 5 | December 31, 2025 | None offered | Customer-specific maintenance only | Rimini Street |
| Oracle E-Business Suite 12.2, JD Edwards EnterpriseOne, PeopleSoft | At least 2037 (Premier Support) | Not yet announced | Full support, including new functionality through continuous innovation releases | ERP Today, ERP Research |
| Oracle E-Business Suite 12.1 | Ended December 2021 (Premier) | December 2025 (Extended) | Sustaining Support: no new patches, no regulatory updates | ERP Research |
| Microsoft Dynamics GP | End of 2029 (technical support and regulatory updates) | Security patches into 2031 | Security updates only after 2029 | ERP Today |
| Microsoft Dynamics AX 2012 R3 | Ended January 10, 2023 | None | Nothing. No security patches since 2023 | endoflife.ai |
Two cautions on this table. First, verify each date on the vendor's own lifecycle page before you put it in a board paper; the sources above are reliable, but vendors change terms, and the SAP 2033 option in particular has conditions. Second, "mainstream support ends" does not mean the software stops. Every system in this table will keep running on January 1 of the year after its date. What changes is what the vendor will do for you, and at what price.
Support ends in phases, and the regulatory phase forces action
Vendors retire support in phases, and the phases matter more than the headline date. The first things to go are new features and performance improvements. Then legal and regulatory updates stop, which is the point at which a finance team starts to feel it: a tax rate change or a new reporting rule has to be built by hand or bought from a third party. Security patches usually continue into the paid extension window and stop when it closes. Bug fixes for problems the vendor already knows about tend to last longest.
For most companies the regulatory line is the one that forces action. An ERP that cannot take a payroll tax update is a compliance problem within a year, whatever the vendor's support page says. Security patches are the second line, and the one the audit committee will ask about. In the SAP case, both lines hold through 2030 for customers who pay for extended maintenance. In the Oracle case, both hold for more than a decade. For Dynamics AX, both lines were crossed in 2023.
SAP has three dates, not one
SAP customers face a sequence, and most of the content written about it treats the sequence as a single cliff. Mainstream maintenance for ECC on enhancement packs 6 through 8 ends on December 31, 2027. From 2028 to the end of 2030, SAP sells extended maintenance at a premium of two percentage points on the license value, which is roughly a 9% increase in the maintenance bill, since standard maintenance runs at about 22% of license value (Rimini Street, ERP Research). Redress Compliance, a license negotiation firm, reports quotes of two to four points before negotiation (Redress). SAP has since added a third option: a private cloud subscription that carries ECC support to 2033 for customers who commit before 2028 (The Register).
The money looks like this on three maintenance bases, using the two-point premium and before any negotiated relief.
| Annual SAP maintenance today | Extended maintenance premium per year | Premium over 2028 to 2030 |
|---|---|---|
| $5M | About $450,000 | About $1.4M |
| $10M | About $900,000 | About $2.7M |
| $20M | About $1.8M | About $5.4M |
The premium is small next to the migration itself. Redress Compliance, a license negotiation firm, puts a large S/4HANA migration at $5M to $50M (Redress). The risk in extended maintenance is the exit terms, not the price. Once a customer is billed for it, there is no mechanism to leave early; a company that commits through 2030 pays through 2030 even if it goes live in 2029 (Redress). That is a negotiation point to settle before signing, not after.
The scale of the backlog is the other number worth knowing. Gartner counted roughly 35,000 ECC customers, and by the end of 2024 about 39% of them had bought or subscribed to S/4HANA transition licenses (The Register). Buying a license is not the same as being live. On the 36-month rule below, a company that had not started by the end of 2024 cannot be live by the end of 2027 without compressing the program. Compression is where overruns come from.
Oracle has no support deadline before 2037
Oracle's on-premise ERP products have no near deadline. In March 2026 Oracle confirmed Premier Support for E-Business Suite 12.2 through at least 2037, and the same commitment covers JD Edwards EnterpriseOne and PeopleSoft under its Applications Unlimited program (ERP Today). Premier Support is Oracle's full tier: new functionality, regulatory updates, and security patches. The only Oracle date that has bitten recently is EBS 12.1, which left Extended Support in December 2025 and now sits on Sustaining Support with no new patches (ERP Research). Customers still on 12.1 have a version problem, not a product problem, and the fix is an upgrade to 12.2, not a platform change.
The "EBS is dead" line has circulated for years, and ERP Research's correction is blunt: EBS is a mature, fully supported product, and the narrative drives poor decisions (ERP Research). The claim is common enough in partner marketing that an Oracle customer should expect to hear it. The right response is to ask for Oracle's lifecycle policy document.
For an Oracle customer, the move to Oracle Fusion Cloud ERP or to another platform is a business case decision with no external clock. That is a better position to be in than SAP's customers have, and it should be used, not wasted.
Microsoft has one real date and one already passed
Dynamics GP loses technical support and regulatory updates at the end of 2029, with security patches continuing into 2031 (ERP Today). Microsoft's intended path is Dynamics 365 Business Central, and for most GP customers, who are small and mid-size, that is a tier 2 program of 6 to 18 months rather than a tier 1 one (Sysgenpro, an SI's published range). The 2029 date is therefore manageable for a company that decides in 2027.
Dynamics AX is the quieter problem. Every AX version is past end of life; the last one, AX 2012 R3, lost extended support on January 10, 2023 (endoflife.ai). Companies still running AX have had no vendor security patches for almost four years. An AX customer has already passed the decision date. The open question is how to explain the current exposure to the audit committee while the program runs. The natural path is Dynamics 365 Finance and Supply Chain, which is a tier 1 program in scope and timeline. Some AX customers use the moment to look at Infor, IFS, Epicor, or Oracle NetSuite instead.
Tier 2 vendors have release dates, not a cliff
Infor, IFS, Epicor, NetSuite, Acumatica, and QAD mostly sell cloud subscriptions with continuous releases, so there is no single support cliff to plan around. The equivalent risk is version currency on older on-premise products (Infor LN and M3 on-premise, Epicor 9 and 10, IFS Applications 9 and 10), where the vendor's support for a specific release does lapse. Customers on those releases face the same planning question SAP customers face, on a smaller scale. They need to know when their release stops getting regulatory and security updates, and whether they can upgrade in place or need a reimplementation.
Consider a $300M distributor on Infor LN on-premise, two releases behind. The vendor's support for its release ends in 18 months, an in-place upgrade is quoted at nine months, and a move to Infor CloudSuite at 14. The decision is the same shape as the SAP one, with smaller numbers, and it should be made the same way, by working back from the support date.
The 36-month rule
A tier 1 ERP program needs 36 months from the decision to proceed to a stable go-live. Published ranges put large implementations at 12 to 36 months (ERP Research), and the short end of that range describes the build, not the program. Selection and contracting take six to nine months on their own. Data cleanup starts before the SI arrives and runs past go-live. A global or multi-entity rollout adds waves. In the programs we have advised, 36 months is the planning number, and companies that budget less either compress testing or go live in waves they did not plan for.
Applied to the dates above, the rule produces a calendar most boards have not seen written down.
| If your system is | And the date that matters is | Then the decision needed to be made by | Position today (November 2026) |
|---|---|---|---|
| SAP ECC, aiming to be off ECC at end of mainstream maintenance | December 31, 2027 | End of 2024 | Late. Extended maintenance or a transition option is now part of the plan, not a fallback |
| SAP ECC, using extended maintenance | December 31, 2030 | End of 2027 | On time if the program starts within the next 12 months |
| SAP ECC, using the private cloud transition option | 2033 | 2030, with the subscription signed before 2028 | Comfortable, but the subscription itself is a commitment to sign within 14 months |
| Dynamics GP (tier 2 path to Business Central, 6 to 18 months) | End of 2029 | Mid-2028 | Comfortable |
| Dynamics AX | Already past | Already past | Exposed. The program should already be running |
| Oracle EBS 12.2, JD Edwards, PeopleSoft | At least 2037 | 2034 | No external clock. The business case sets the date |
Regulated utilities should add six to twelve months to the 36. A utility that wants to recover the program's cost through a rate case has to time go-live so the new system's costs land in the test year the filing is built on, and commissions and auditors ask for a clean cutover before they accept the numbers. Consider a utility planning a 2030 go-live with a rate case filed in early 2031 on a 2030 test year: the program has to be stable by the third quarter of 2030, which means the decision date moves from 2027 to the middle of 2026.
Extended support is a bridge, not a strategy
The question SAP customers ask most often is whether extended maintenance or third-party support can carry them for a long time instead of a program. In our experience, no company uses either for an extended period. Every one we have seen used it as a bridge to a planned go-live, and the ones that used it well decided to use it early, built the premium into the program budget, and negotiated the terms (especially the exit terms) while they still had bargaining power.
The arithmetic supports that pattern. Third-party support providers price at roughly half the vendor's maintenance bill, according to Redress Compliance, a license negotiation firm; Rimini Street, the largest such provider, claims savings of up to 90% (Redress, The Stack). Both are sellers of the service, so the number to use in a business case is the quote you receive, not either claim.
What a third-party contract does not include is regulatory content from the vendor, new functionality, or a path back to vendor support without a reinstatement fee. For a company with a go-live date two or three years out, those gaps are tolerable. For a company with no date, they compound every year. The eventual program gets harder, not easier, because the gap between the old system and the new one keeps widening.
The decision rule, then, is a function of one number: where your realistic go-live lands.
- If your go-live lands inside the vendor's paid extension window (for SAP, before the end of 2030), buy the extension. It keeps regulatory and security coverage from the vendor during cutover, and the premium is small next to the program.
- If your go-live lands beyond the extension window, bridge with third-party support or the vendor's transition subscription, and treat the bridge as a line in the program budget with an end date.
- If your vendor has no near deadline (Oracle, most tier 2 cloud products), the support date is not your trigger. Decide on business value, and do not let anyone put a clock on the slide that the vendor has not put on its lifecycle page.
Vendor incentives follow the fiscal quarter, not the support date
It is widely assumed that vendors use support deadlines to pressure customers into migrating. We have not seen it work that way. In negotiations we have been part of, the SAP, Oracle, and Microsoft account teams did not lean on the support date. What they did offer were financial incentives to commit faster, and those were tied to the vendor's quarter-end and year-end close rather than to any customer deadline.
That changes the advice. The support date tells you when you must decide; the vendor's fiscal calendar tells you when to sign. SAP's fiscal year ends in December, Oracle's in May, and Microsoft's in June.
A customer who is ready to commit in the last six weeks of a vendor's quarter, and especially its fourth quarter, will see better terms than one who signs in the first month. The difference shows up in subscription pricing, transition credits, and extended maintenance relief. The practical sequence is to finish selection and internal approval on your own timeline, then hold the signature for the vendor's quarter-end. The converse also holds: a vendor's "this offer expires" message in the last week of its quarter is about the vendor's number, and it will usually reappear next quarter.
The incentives themselves take a few familiar shapes. A transition credit counts part of the existing license value against the new subscription. A maintenance holiday waives the premium for the first year of an extension. A subscription ramp starts the new fee at a fraction of the full rate and steps it up over two or three years.
Suppose a customer with a $10M maintenance bill is offered a 40% first-year discount on the subscription and a credit worth 15% of the subscription if it signs in December rather than February. On a $12M annual subscription that is $7.2M in year one instead of $12M, plus $1.8M in credits, roughly $6.6M of difference for a ten-week change in signing date. That is why the signature waits for the vendor's quarter, and the selection does not.
Run your own dates
The deadline planner takes your current system, your annual maintenance bill, and your realistic go-live and returns the five-year cost of migrating now, extending, or bridging, side by side.